# DeepSeek-V2.5 (Sep 2024) vs Llama 3.1-70B

> DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 29.4 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v2-5-vs-llama-3-1-70b
- Last updated: 2026-10-10
- Shared benchmarks: 19

## Summary

- They share 19 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 8 categories and Llama 3.1-70B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 33.3 to 12.7.

## Snapshot

| | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 37.0 | 29.4 |
| Rank | 197 | 299 |
| Context | — | 128K |
| Input $/M | — | $0.40 |
| Output $/M | — | $0.40 |
| Weights | Open | Open |

## Coding

- DeepSeek-V2.5 (Sep 2024): 31.3 (#274)
- Llama 3.1-70B: 29.8 (#292)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| BigCodeBench Instruct | 48.6% | 46.1% |
| LMArena Coding | 1310 | 1260 |
| BigCodeBench Complete | 53.2% | 54.8% |
| Aider Polyglot | 17.8% | — |
| WeirdML | — | 9% |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |

## Agentic & Tool Use

- DeepSeek-V2.5 (Sep 2024): —
- Llama 3.1-70B: 24.9 (#104)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| TheAgentCompany | — | 6.9% |
| BALROG | — | 27.9% |

## Reasoning

- DeepSeek-V2.5 (Sep 2024): 24.5 (#145)
- Llama 3.1-70B: 20.8 (#215)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1241 |
| DTBench | — | 60% |
| LMCA | — | 14.8% |
| Epoch Capabilities Index | — | 125.91 |

## Math

- DeepSeek-V2.5 (Sep 2024): 33.3 (#177)
- Llama 3.1-70B: 12.7 (#298)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Math | 1289 | 1252 |
| OTIS Mock AIME 2024-2025 | — | 3.6% |
| Omni-MATH | — | 21% |
| MATH Level 5 | — | 36.7% |

## Knowledge

- DeepSeek-V2.5 (Sep 2024): 34.2 (#190)
- Llama 3.1-70B: 24.1 (#264)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Expert | 1266 | 1209 |
| GPQA Diamond | — | 44.2% |
| MMLU-Pro | — | 65.3% |
| GPQA (HELM) | — | 42.6% |
| MMLU | — | 80.1% |

## Multilingual

- DeepSeek-V2.5 (Sep 2024): 42.7 (#188)
- Llama 3.1-70B: 38.9 (#220)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Non-English | 1273 | 1219 |
| LMArena Chinese | 1317 | 1215 |
| LMArena French | 1288 | 1260 |
| LMArena German | 1258 | 1222 |
| LMArena Japanese | 1228 | 1132 |
| LMArena Korean | 1209 | 1140 |
| LMArena Russian | 1289 | 1234 |
| LMArena Spanish | 1248 | 1253 |

## Instruction Following

- DeepSeek-V2.5 (Sep 2024): 67.7 (#189)
- Llama 3.1-70B: 65.5 (#218)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1231 |
| IFEval | — | 82.1% |

## Long Context

- DeepSeek-V2.5 (Sep 2024): 39.5 (#169)
- Llama 3.1-70B: 37.6 (#208)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Longer Query | 1301 | 1241 |

## Writing & Preference

- DeepSeek-V2.5 (Sep 2024): 50.0 (#182)
- Llama 3.1-70B: 35.6 (#262)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-70B |
|---|---|---|
| LMArena Text | 1294 | 1261 |
| LMArena Creative Writing | 1285 | 1232 |
| LMArena Multi-Turn | 1297 | 1256 |
| EQ-Bench Creative Writing | — | 784 |
| WildBench | — | 75.8% |

## FAQ

### Is DeepSeek-V2.5 (Sep 2024) better than Llama 3.1-70B?

DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 29.4 on the Noometry Index.

### Is DeepSeek-V2.5 (Sep 2024) or Llama 3.1-70B better for coding?

DeepSeek-V2.5 (Sep 2024) scores higher on coding benchmarks: 31.3 versus 29.8 in the Noometry coding category.

### How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Llama 3.1-70B share?

19 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Llama 3.1-70B has 35.